基于高维数据分组的陶瓷产品需求预测算法  

PREDICATION OF CERAMIC PRODUCT DEMAND BASED ON HIGH DIMENSIONAL DATA PARTITION

作　　者：詹棠森[1] 刘伟洁[1] 

机构地区：[1]景德镇陶瓷学院信息工程学院,景德镇333403

出　　处：《陶瓷学报》2009年第2期239-242,共4页Journal of Ceramics

基　　金：赣财教[2008]147号省科技厅项目

摘　　要：提出了一种高维数据的最优层次分析组合预测算法,即通过相关系数得到影响因素的分组,并得到相应的权值,通过设定参数权值向量和最优非线性问题,得到比较好的预测值,比没有分组得到的结果更好,这样克服了主观定义成对比较矩阵及一致性检验,从而具有更广泛的实用性。A combined predication algorithm was produced based on optimal AHP of high dimensional data, by which the partition of influencing factors and weight values were obtained through correlation coefficient calculation. Better predication results were gained with the determination of parameter weight value vector and optimal nonlinear problems. The prediction algorithm can overcome the subjective definition in paired comparison matrix and consistency test, and is capable of wide application.

关 键 词：分组 参数权值向量 最优非线性问题 

分 类 号：TQ174[化学工程—陶瓷工业]